Top
Best
New

Posted by crorella 9 hours ago

GPT 6.1 Sol: Near-Astra intelligence for a fifth of the price(openai.com)
799 points | 737 commentspage 7
AmazingTurtle 7 hours ago|
regarding the new pro max subscription btw:

it's 500$ for 25x the plus usage, thats pro (max).

this implies that the old 200$ 20x pro (more) is now more like 10x the usage of plus.

they are slashing our subscriptions in half and make it "but we're more efficient!"

user43928 6 hours ago|
Which they are, so another way to see it would be that they make the API/enterprise cheaper.

However, I don't know how future larger models such as the cancelled 6.1 Astra will be priced.

If the price stays high, this would indeed be quite bad for the $200 subscription..

dangoodmanUT 8 hours ago||
I can appreciate they show that Opus 5.5 is objectively better intel. and cost wise on multiple benchmarks
m4rtink 8 hours ago||
Price wars, those always end well, especially if you are loosing billions per day!
demibabs 8 hours ago||
Is OpenAI’s Cloudflare turnstile new? Never gotten it before.

Either way, a little ironic…

unsupp0rted 8 hours ago||
Excited for Gemini 4 at equal or better coding and a fifth of the price of 6.1 Sol
oh_no 8 hours ago|
3.8 flash is more expensive than sol 6.0, google uses a LOT of reasoning tokens
Lapalux 9 hours ago||
It's relentless, isn't it?
nilslindemann 6 hours ago||
Thanks, but I need the money for the next laptop.
zf00002 9 hours ago||
I can blow through my weekly on astra in a few hours; hopefully this really is as good.
skybrian 9 hours ago|
I get decent results telling it to use Luna subagents for implementing commits.
MisterMunchkin 9 hours ago|
Still too expensive. Needs to come down to $1 or less to compete with china.
AspireOne 4 hours ago|
It did. And 6.1 now outcompetes China in most cases.

Chinese models are cheap and fast by token, but they generate oceans of thinking tokens in order to accomplish the same result GPT-6.1-Sol accomplishes in, comparatively, two drops of thinking tokens, making the Chinese models come out costlier and slower per actual task performed end-to-end.

Here's from artificialanalysis.ai, per task:

* Deepseek-v4.1-flash (max): 0.27$, 5.5 minutes, 89k tokens generated.

* GPT-6.1-Sol (medium) : 0.21$, 2.2 minutes, 8k tokens generated.

API pricing. With Sol having SUBSTANTIALLY better performance.

More comments...